Development and Validation of an Interpretable Machine Learning Model for Predicting Venous Thromboembolism（VTE）in Intensive Care Unit (ICU) Patients

Development and Validation of an Interpretable Machine Learning Model for Predicting Venous Thromboembolism（VTE）in Intensive Care Unit (ICU) Patients

Venous thromboembolism remains a leading cause of preventable mortality in intensive care unit (ICU) patients. Existing risk-stratification tools were developed in general medical populations and lack ICU-specific predictors. This study was to develop and validate an interpretable machine learning (ML) model to predict VTE in ICU patients.

Inclusion Criteria:

• age ≥18 years;
• ICU length of stay ≥48 hour
• the first ICU admission

Exclusion Criteria:

• VTE diagnosed prior to ICU admission
• VTE diagnosed within 24 hours of ICU admission
• >20% missing values in key variables